Agricultural development and the distribution of water resources in Kgatleng District, Botswana

Abstract

During Botswana's four decades of high levels of growth the agricultural sector has lagged behind, with smallholder productivity being especially poor. This paper applies an equity perspective: its main claim is that one important explanation for the current lack of agricultural development is the unequal distribution of agricultural resources. It takes into account both the national institutional structure, which promotes widespread inequality, and the distribution of boreholes and water resources on the communal grazing range in Kgatleng District. It argues that ever since the first administrative effort to develop water resources in the 1920s the country's official policy and legislation has directly or indirectly favoured the large-scale farmers over the smallholders and, further, that customary property rights principles have supported the process that has led to today's institutional inequality.

Keywords:

• agricultural development
• equity
• water
• Botswana

1. Introduction

Despite the growth miracle of Botswana's last four decades, there are concerns about the lack of structural transformation in the country's economy (Hillbom, 2008). Of particular concern are the low levels of productivity in the agricultural sector and the high levels of national inequality (see e.g. Mpabanga, 1997; Leith, 1997, 2005; Hillbom, 2008). The main arguments in this article rest on a conviction that a country's structural change and economic development depend on productivity increase in the agricultural sector. This is in line with the writings of economists such as Lewis (1955), Kutznets (1973) and Mellor (1986). In the existing equity literature, poor agricultural performance is also causally linked to inequality. Unequal distribution of agricultural resources is considered to be embedded in structures of inequality at the national level and this hampers agricultural development (see e.g. Cornia, 1985; Deininger & Squire, 1996; Engerman & Sokoloff, 2002; Griffin et al., 2002). The call for equity is therefore more than just a question of fairness and concern for people's well-being; it is also a question of optimal allocation and use of resources in order to achieve economic growth and development. Equity and distribution of resources are usually measured in terms of incomes and land, and empirical investigations have been conducted primarily in Latin America and Pacific Asia (see e.g. Kay, 2001; Engerman & Sokoloff, 2002; Griffin et al., 2002). This article contributes to the debate by expanding it to the African context, specifically investigating the distribution of water, a key agricultural resource, in the cattle sector.

One explanatory factor in the unequal distribution of agricultural resources that is often debated is the impact of colonialism and the expropriation of natural resources by colonial administrations. In Africa, colonial involvement varied according to local economic opportunities and political interests. In former settler colonies such as South Africa, Zimbabwe and Kenya, the colonial administration expropriated land from the African population while at the same time keeping wages low and restricting the movement of the African labour force. The goal was to protect the economic interests of the European farmers and supply them with cheap land and labour (see e.g. Austen, 1987:168–71; Berry, 1993:101). In Botswana, however, the colonial influence was limited in comparison with that of other countries in Africa and abroad. While the British colonial administration's headquarters were located in neighbouring South Africa, the Tswana dikgosi (plural of kgosi, chief) enjoyed a high degree of independence. Both the dikgosi and the British saw the Protectorate as a primarily African territory where settlement by white farmers and European-style development were discouraged (Parsons & Crowder, 1988:xvii–xix). Nonetheless, property rights allocating agricultural resources were indirectly affected by the colonial policies in the Protectorate. As in the rest of colonial Africa, the administration was involved in formulating customary law and laws of tenure (see e.g. Chanock, 1991; Berry, 1993; Mamdani, 1996). The establishment of customary property rights became a joint venture of the colonial administration and the tribal leaders. Since customary property rights institutions prevail in current legislation and agricultural policy, it is necessary to analyse them if we are to explain today's unequal distribution of agricultural resources.

The article has three aims: to highlight the importance of equal distribution of water resources by analysing Botswana's poor productivity at the national level in the cattle sector and its weak agricultural development generally; to trace the origin of the present institutional inequality in Kgatleng District, as evident in the principles guiding its legislation and policy, back to customary law and practices governing water; and to connect the existing property rights institutions at the district level to the national institutional structure and explain the persistence of structural inequality.

The primary material for this article was collected during two fieldwork periods in Botswana. In 1999/2000 archive studies were conducted in Gaborone and interviews with water users were carried out in rural Kgatleng. In 2007 a follow-up study was conducted to revisit water users and repeat interviews.

2. Institutional inequality

In four decades Botswana has gone from being one of the poorest countries in the world to its current position as a high-middle-income country (World Bank, 2008). The country is also politically stable and has one of the longest running multi-party systems in sub-Saharan Africa. Mainstream explanations for the ‘Botswana success story’ include limited colonial influence, good political institutions and wise leaders (see e.g. Samatar, 1999; Acemoglu et al., 2003; Leith, 2005; Beaulier & Subrick, 2006; Iimi, 2006). Nonetheless, some researchers point to serious deficiencies in the country's politico- and socio-economic structure, such as elite capturing, widespread and persistent inequalities, poor and neglected rural areas, low agricultural productivity, high unemployment rates, failure to limit the AIDS epidemic, discrimination against minority groups such as the San, and a weak civil society (see e.g. Good, 1992, 1993, 1994; Gulbrandsen, 1996; Nthomang, 2004; Wikan, 2004; Heald, 2005; Lekoko & Van der Merwe, 2006; Makgala, 2006; Phaladze & Tlou, 2006; Hillbom, 2008). A growth miracle that is primarily dependent on earnings from diamond exports cannot by itself prove the overall existence of good governance and a development-friendly institutional structure. The evidence of dual development and the prevalence of contradictory indicators necessitate further evaluation of the institutional structure, and separately from the growth performance. This includes investigating property rights structures that guide the allocation of agricultural resources such as water.

With a national GINI coefficient of 0.6 (UNDP, 2008), Botswana's income differences are among the highest in the world. Despite both a trickle down from the wealthier segments of society and a political consciousness of the need to fight poverty, roughly 30 per cent of the population is currently living below the national poverty line (UNDP, 2005:6). Further, there have for decades been reports of prevailing and even increasing rural poverty and a polarisation of agricultural resources (see e.g. Parson, 1981; Good, 1993, 1994; Peters, 1994; Gulbrandsen, 1996; Wikan, 2004). Meanwhile, government policy since independence has mainly been the equivalent of creating a basic needs system. Health care, education and support have been provided to the poor via food-for-work programmes (Colcough & McCarthy, 1980; Peters, 1994; Gulbrandsen, 1996; Nthomang, 2004; Lekoko & Van der Merwe, 2006). This is not, however, the same as having a policy for targeting redistribution of wealth and resources. This article takes the debate on poverty in Botswana further by suggesting that inequality is the root cause of the rural areas' generally poor agricultural performance.

The call for a more even distribution of natural resources in the agricultural sector is prompted by the argument that there is a causal relationship between equity and agricultural development and growth which would improve incomes and living standards for the rural population (see e.g. Berry & Cline, 1979; Cornia, 1985; Deininger & Squire, 1996; Kay, 2001; Griffin et al., 2002). The larger the holding of the resource, the lower the value per unit for the property holder, who therefore attempts to increase output per labour unit, as labour in this case is the scarce production factor. Meanwhile, the situation is the opposite for the smallholder who relies on family labour and on increasing output per resource unit. As the relative factor price of agricultural resources increases, the resource holder will intensify farming methods, which in turn leads to an increase in land productivity and a decrease in labour productivity. In societies with abundant labour and scarce agricultural resources and capital, total factor productivity should increase after redistributive reform since smallholders on the whole use resources more efficiently than larger land holders (Cornia, 1985; Deininger & Squire, 1996; Griffin et al., 2002:185–7).

An unequal distribution of agricultural resources not only hinders agricultural development but, more seriously, reflects an institutional inequality that characterises society at large (Engerman & Sokoloff, 2002; World Bank, 2005). In societies with institutional inequality, government policies tend to favour elites in various ways while denying the majority of the population equal opportunities. Groups and individuals with wealth and high social status influence institutions and distort them in their favour, further enriching themselves. Political and economic inequality thus go hand in hand and become institutionalised, perpetuating these inequalities (World Bank, 2005:107–8).

The equity literature referred to above is mostly based on empirical evidence from crop farming in Asia and Latin America. Some studies, however, suggest that the causal relationships identified in those countries are also at play in cattle farming in sub-Saharan Africa. Smith's (2004) study of smallholders on communal land in Zambia reveals a significant positive causal correlation between secure land tenure, fixed investments and productivity increase in cattle husbandry; data collected in South Africa by Mapiye et al. (2009) indicates that herds on the communal grazing range that have access to more natural resources also have a higher productivity rate; and a study by Malley et al. (2009) in Tanzania shows that pro-poor rural policies in agricultural development could lead to improved productivity. Growing evidence demonstrates that the equity argument is relevant for the African context and for the cattle sector.

Botswana has a profile of abundant labour and scarce agricultural resources. Roughly 40 per cent of its population live in the rural areas (World Bank, 2008). With unemployment at 15.8 per cent (UNDP, 2008), labour productivity could be increased if smallholders controlled more resources. As only 5 per cent of the land is arable, livestock is the backbone of the country's agricultural sector. Although the carrying capacity of the grazing range has been exceeded in some areas, there is generally enough land for farmers in the customary areas. The primary restricting factor, for crop production as well as for smallstock and cattle, is access to permanent water sources (see e.g. Thirtle et al., 2003; Emongor, 2006). Consequently, water, not land, is the country's most scarce and valuable agricultural resource.

Most Batswana farmers today are agro-pastoral subsistence smallholders. During the 40 years of independence and significant economic growth these farmers have seen little increase in productivity. Fields and animals often yield even less than the requirements for subsistence and many households depend on complementary off-farm incomes (Arntzen & Silitshena, 1989; Gulbrandsen, 1996:2–3). The smallholder sector has made no visible contribution to the present national economic growth and has experienced neither technological nor institutional change. Its low rates of productivity are exemplified by the off-take rates among traditional small-scale cattle holders; that is, those who keep their cattle on the communal grazing range (see Table 1). Cattle of course bring other economic benefits to the farmer, for example as draught animals and for milk, but these uses are of less importance for the argument being made here, since they are neither part of large commercial networks nor do they generate export income as in the case of meat. There are also social benefits from cattle, such as bride wealth and maintaining social ties, but these also fall outside the scope of this article.

Table 1: Cattle holdings and off-take rates in the traditional sector

District/region	Cattle population		Cattle/holding^a	Off-take	Off-take rates^b
	Number	Percentage
Barolong	23 379	0.9	22	2613	11.2
Ngwaketse South	68 128	2.8	27	4090	6.0
Ngwaketse North	58 511	2.4	15	2295	3.9
Ngewaketse Central	106 795	4.3	20	9081	8.5
Ngewaketse West	71 370	2.9	30	5817	8.2
Southern region total	328 183	13.3	23	23 896	7.3
Balete/Tlokweng	2222	0.1	10	161	7.2
Kweneng South	89 238	3.6	22	8970	10.1
Kweneng North	78 584	3.2	28	7564	9.6
Kweneng West	209 627	8.5	44	16 785	8.0
Kgatleng	110 026	4.5	21	6375	5.8
Gaborone region total	489 697	19.8	25	39 855	8.1
Mahalapye East	42 497	1.7	19	2079	4.9
Mahylapye West	579 013	23.5	101	32 248	5.6
Palapye	104 048	4.2	21	4719	4.5
Serowe	110 617	4.5	51	6644	6.0
Bobonong	81 606	3.3	27	14 142	17.3
Letlhakane	221 737	9.0	80	16 212	7.3
Selebi-Phikwe	22 548	0.9	43	2554	11.3
Central region total	1 162 066	47.1	49	78 598	6.8
Tati	6091	0.2	11	384	6.3
Tutume	95 724	3.9	36	11 659	12.2
Totona	30 919	1.3	31	3246	10.5
Francistown region total	132 734	5.4	26	15 289	11.5
Ngamiland West	109 186	4.4	43	4517	4.1
Ngamiland East	64 288	2.6	41	6982	10.9
Chobe	9433	0.4	20	658	7.0
Maun region total	182 907	7.4	35	12 157	6.6
Gantsi	72 276	2.9	62	7219	10.0
Hukuntsi	15 309	0.6	26	1890	12.3
Tsabong	85 227	3.5	57	10 379	12.2
Western region total	172 812	7.0	48	19 488	11.3
Total	2 468 399	100	34	189 283	7.7
^aAverage number of head of cattle per cattle holder. ^bPercentage of the total stock that is sold or slaughtered.
Source: Adaptation of RoB (2003:37, Table 4.1, 43, Table 4.6).

Table 1 shows generally low off-take rates, ranging from 3.9 per cent to 17.3 per cent. The average for the country is 7.7 per cent and Kgatleng falls below this at 5.8 per cent. To put these figures into perspective, Table 2 shows international off-take rates, demonstrating that in comparison with other developing countries, including neighbouring South Africa, Botswana is performing poorly. Even the high performance of Bobonong District falls short when compared with international rates.

Table 2: International off-take rates compared with Botswana

Country	Off-take rates^a
	1967 (%)	1977 (%)	1987 (%)	1997 (%)	2007 (%)
Argentina	24	24	25	26	27
Australia	31	38	36	31	32
Botswana	9	6	8	9	7
Mongolia	19	19	21	20	19
South Africa	21	21	23	16	22
^aPercentage of animals slaughtered out of the national stock.
Source: FAOSTAT (2009).

Despite the declining value of livestock and the degradation of the grazing range, wealth and poverty among Batswana farmers still essentially take the form of ownership of cattle. Consequently, differences in disposable income and increasing rural poverty are mirrored by the increased number of smallholders without cattle and a polarisation in cattle ownership. In the 1940s Isaac Schapera found that only 10 per cent of households held no livestock (Hesselberg, 1985:182). In the first decade after independence the national herd expanded significantly, but at the same time many rural households lost their livestock. An agricultural survey in 1983 reported that 30 per cent of all rural households owned no cattle, while 30 per cent of the national herd was owned by 4 per cent of households (Arntzen & Silitshena, 1989:162). In the early 1990s roughly half of rural households held no cattle, while 5 per cent owned 50 per cent of the national herd (Gulbrandsen, 1996:3). Table 3 provides further evidence that the distribution of cattle has been continuously polarised during more than half a century. There are no indications that this trend has been reversed.

Table 3: Distribution of cattle ownership in Botswana, 1981–1995

Farming households	1981 (%)	1990 (%)	1993 (%)	1995 (%)
No cattle	32	38	47	49
1–40 head	41	38	36	25
40–100 head	18	18	11	16
>100 head	9	6	6	10
Total	100	100	100	100
Source: Wikan (2001:31, Table 3).

For smallholders to hold cattle, water rights are crucial. Crop farming in Botswana is generally rainfed, but to support livestock it is necessary to hold property rights to permanent water sources, such as boreholes and wells on the grazing range. Difficulties with accessing and remaining in control of these sources can force smallholders to sell off their cattle, and farmers who for reasons other than lack of access to water resources have lost their cattle have little or no opportunity to keep their water rights. To retain property rights requires an active use of the rights. The farmer whose animals are gone and who no longer uses the water point will lose the basis for the right and therefore easily be excluded and become merely a hirer or dependant, to the benefit of someone else. This is discussed further in Sections 3 and 5 below. Staying on as a syndicate member without having cattle is expensive as the former cattle holder may be obliged to continue contributing towards the running of the borehole (interviews with syndicate members, permanent hirers and dependants, Kgatleng District, June 2000 and March 2007). When smallholders lose their cattle, water rights are transferred to larger cattle holders who still have animals to water. The polarisation of cattle ownership is thus causally related to an increasingly unequal distribution of water rights (Peters, 1994).

3. Negotiating customary property rights

Extensive anthropological investigations into the laws and customs of the Tswana were conducted in the 1930s and 1940s. The work of Isaac Schapera (e.g. Schapera 1980, 1994) marked the start of an increased colonial awareness and registration of the local legislation in the Bechuanaland Protectorate. What was then registered as Tswana Customary Law has to a large extent been preserved in the current legislation. There is some debate about the political context in which the colonial administration and the indigenous elite, the chiefs, operated and how they came to formalise ‘customary’ law and tenure in Africa. It has been argued that this formalisation allowed expropriating tribal leaders to profit by amassing resources (see e.g. Chanock, 1991; Mamdani, 1996).

Generalising about the sub-Saharan African customary property rights systems that govern natural resources is problematic, given the sub-continent's wide range of historical experiences, political systems, environments and cultures. Nevertheless, it is possible to detect some general guiding principles that are also applicable to water rights in Botswana in general and Kgatleng in particular. Natural resources were, as an overriding principle, seen as belonging to the community in its entirety and kept in trust by its leaders, but there was wide variation in who controlled and was allowed to access the resources in practice. Resource allocations were made by the chief together with specific community leaders, and community memberships ensured that primarily male heads of household had private rights to access and exploit land, water and other resources. While individuals were allocated user rights, the community retained the right to decide the rules governing exploitation and to exclude rule breakers. By applying individual labour to communally held resources, farmers obtained private ownership of structures such as water points, and of their own agricultural production. Those who did not use the resource actively could have their rights withdrawn (see e.g. Perrings, 1992; Berry, 1993; Ensminger, 1997). Investments in natural resources have, consequently, been an important strategy for claiming a resource and obtaining security in property rights (Sjaalstad & Bromley, 1997). Sub-Saharan African customary property rights are flexible and negotiable, and the ability to negotiate and secure access to resources depends on a person's social, political and economic status. Most researchers agree that this negotiability ultimately protects and promotes the opportunities of those who have a low social and economic status (see e.g. Berry, 2002; Odgaard, 2003). Others, however, claim that since stakeholders are not socially, politically and economically equal, negotiations are used by the more powerful to enrich themselves (see e.g. Peters, 1994, 2004).

In Kgatleng, natural surface and ground waters are regarded as a gift from nature to those who are able to make use of them, and membership in the local community gives people access to them. As no individual can claim to have constructed natural waters, they are treated as a public good, a classification which is also linked to the life-supporting qualities of water. Property rights to constructed water sources such as boreholes, on the other hand, are guided by principles similar to those that guarantee a person private rights to agricultural production. Labour and financial inputs in construction and maintenance together make up the foundation for property rights claims. These principles allow the wealthier farmers to invest and thereby gain private or exclusive communal property rights to water points, while the poorer farmers, who are members of larger groups pooling their resources, have to settle for less exclusive property rights (Peters, 1994; Carlsson, 2003).

Looking at property rights to water in Kgatleng in a broader social, political, and economic context raises questions about negotiability and how it can increase exclusion. Today's structural inequalities and polarisation of agricultural resources in Botswana are rooted in the pre-colonial and colonial tribal redistributive system, which allowed an elite privileged access to agricultural resources and cattle (Wylie, 1990; Peters, 1994; Good, 1994; Gulbrandsen, 1996; Carlsson, 2003). This situation, where the present elite support and maintain the customary institutional system, is not unique to Botswana but a general trend in sub-Saharan Africa that has the effect of increasing inequality and polarisation (Platteau, 1996; Carney & Farrington, 1998; Ribot, 2000).

4. Borehole development and legislation

Cattle herding and the control and ownership of cattle have long been of great importance to the Tswana for their economic as well as social status, and Tswana leaders early on had an interest in acquiring modern technologies that could improve their production. One such technology, the drilling of significantly deeper boreholes, was introduced in southern Africa in the early 1900s. Boreholes made it easier to water the cattle, protected them in times of drought, and opened up new grazing areas that lacked surface water.

In the late 1920s the BaKgatla1 leaders began drilling in the BaKgatla Reserve. Although only seven of their 16 boreholes were successful, this was a pioneering initiative and the first large-scale water development scheme in Bechuanaland (Schapera, 1980:22–3; Peters, 1994:58). The initiative continued in the 1930s. Some of the tribal boreholes were allocated to individuals of high economic and political status and thus became de facto private property, but most were handed over to ‘borehole syndicates’ – small, exclusive groups usually made up of farmers connected to the tribal elite. The boreholes demanded much larger investments than the earlier hand-dug wells, so property rights holders had to share the costs. By letting a syndicate, rather than the tribe as a whole, bear the costs as well as reap the benefits of the borehole, it was hoped that the syndicate members would realise the long-term gains of maintenance (BNA, 1958–60). From the start there was then a variety in property rights to boreholes – tribal, private and syndicate.

The Tswana dikgosi's demands for water development schemes were eventually supported by the colonial administration. In the 1930s the British Administration's policy started to change and efforts were made to modernise the country and develop natural resources that could benefit export income-earning sectors (Parsons & Crowder, 1988; Steenkamp, 1991). To understand the water development schemes, they must be seen in connection with the strong interest of both colonial and tribal authorities to expand cattle herds and increase incomes from beef exports. The British sought a way to enable the Protectorate to carry its own administrative and development costs, while the tribal leaders for their part wanted to modernise their own reserves and at the same time raise their own personal incomes and those of their political allies. The early water development schemes benefited those who already held substantial herds but not the small-scale cattle farmers (BNA, 1933; interviews with syndicate members in Kgatleng District, June 2000).

In 1940 the British administration sank 27 boreholes in the BaKgatla reserve, 21 of which were successful (BNA, 1961). After this, World War II temporarily brought all drilling to a halt. From 1946 to 1959 another 88 were drilled, with 44 being successful (BNA, 1958–60). The colonial administration continued the tribal policy of handing over the boreholes to individuals and to syndicates of larger cattle holders whose wealth and high status gave them the requisite political influence (Peters, 1994:59–62). The smallholders with few or no cattle demanded that water sources be developed on the arable lands and in residential areas, mostly for domestic use, watering small stock and to a smaller degree for agricultural use, but the colonial administration did not consider their demands to be of economic interest and these farmers therefore had to secure their own water supply.

In the late 1950s and the 1960s the rate of borehole drilling increased. There was renewed interest in economic development in the British colonies beyond a mere continuance of policies that had already been adopted (Peters, 1994:77–8). The administration's concern with water point management persisted. The question of responsibility had not been dealt with satisfactorily, despite the earlier argument that the creation of syndicates would solve this problem. Other schemes were introduced, all with the aim of reimbursing the administration for its drilling and maintenance expenses, but efforts to shift all costs from the public administration to the private beneficiaries failed in one way or another (BNA, 1963/64).

Since to construct a water point on the communal grazing range the first requirement has always been to secure a land right, it is important to understand the system of land allocation in order to understand the control of boreholes. After independence, the state took over from the tribal authorities the responsibility for allocating land to farmers. The first major land law to be enacted was the Tribal Land Act of 1968. With this law the formal responsibility of allocating land in communal areas, including land for boreholes and dams, was vested in Land Boards, which came into existence in 1970 (Gulbrandsen, 1984:8–9). The legislation stated that Land Boards were to allocate land according to the same principles that had been prominent in customary law and that the right received by the farmer would still be a user right for a piece of communal land and not a private property right. Botswana has 12 Land Boards and Kgatleng District has one of its own. The board members are elected and the kgosi, who previously executed all land allocations together with his overseers (appointed by the kgosi to help keep track of local land allocations), is now a non-voting member of the Land Board. Where there are pre-Land Board allocations that have not been brought to the Boards' attention, they have to rely on previous customary authorities (interviews, T Kgatlwane, Department of Lands, Ministry of Government and Lands, Gaborone, 28 April 1999; F Monagong, Secretary, Artesia Sub-Land Board, Artesia, 2 June 2000; K Ramotsebe, Sub-Land Board Clerk, Mamathubudukwane, 29 June 2000).

In 1975 the Tribal Grazing Land Policy was enacted. All land was divided into three categories: communal land that would stay under Land Boards working according to principles of customary tenure, commercial land where groups and individuals would be given exclusive rights in the form of leasehold-tenure, and reserve land that was put aside for future needs and for wildlife. Kgatleng District, like most of the more densely populated eastern part of Botswana, was designated as communal land. It was hoped that the reform would supply incentives for large-scale cattle farmers to move out of communal areas and start up commercial farms (Parson, 1981; Masalila, 1983:152–6; Gulbrandsen, 1984:10–18). Instead, these farmers drilled commercial boreholes in the northwestern parts of Kgatleng District while at the same time making use of communal grazing and water resources. According to government practice, boreholes are spaced 8 kilometres apart, and since cattle only wander about 5 kilometres per day, whoever controls the water points also controls the grazing around them. Consequently, it is deceptive to state that the communal grazing range is open to all; effectively it belongs to the borehole owners. On the basis of ownership of private boreholes, cattle holders have lobbied for private control over the surrounding grazing range despite the communal designation.

The Water Act of 1968 is the first and only statutory water law to be enforced in Botswana. It obliges the Ministry of Mineral Resources and Water Affairs to appoint a Water Apportionment Board, which is the sole government organisation with the mandate to grant water rights. This Act follows the principles of customary law and is mainly concerned with regulating water use for agricultural and commercial purposes. It states that no rights of property can be held to public waters, meaning all surface and ground water that is not used for mining activities (RoB, 1968: Part I:2, Part II:4), and also that no water right is required for watering livestock and taking water for domestic purposes using public waters (RoB, 1968: Part II:5; interviews, SG Sokwane, Land Apportionment Board Secretary, Ministry of Minerals and Water Affairs, Gaborone, 24 and 30 April 1999). If, however, surface waters such as rivers are used for commercial purposes and irrigation schemes, then a water right is required (interview, SG Sokwane, 24 April 1999). No water right is needed prior to constructing a water point for domestic purposes (RoB, 1968: Part II:6), but a land right is needed for the site. If the water source lacks a water right or the water user is extracting more than is stipulated in the water right, there are no serious consequences. Further, no fees are charged for the water consumption (RoB, 1968: Part II:6). All in all, there have been no conscious efforts by the Water Apportionment Board either to address problems of water scarcity or to enforce an equitable distribution of water resources. Water is a de facto public good available to whoever controls the water points.

5. Equity and exclusion among users

African customary property rights systems could be perceived as promoting equality because of the inherent strong resentment of exclusion. The communal ownership of agricultural resources implies a willingness to share with the less fortunate, but there is nevertheless acceptance of the gradual amassing of water rights in the hands of the larger cattle holders that is supported by current customary law. To understand these two parallel views we have to distinguish between the principles of granting basic needs, on the one hand, and striving for equity, on the other. As long as the levels of collective wealth are low the first may in practice resemble the second. As wealth increases, however, the difference becomes clear. In Kgatleng District, competition over rights to permanent water sources has been increasing for decades, especially on the communal grazing range, which has reached its carrying capacity in terms of grazing animals and boreholes, and the toughening competition provokes power struggles (interview, G Tlwane, Senior Chief Representative, Artesia, 28 February 2007).

The borehole syndicates were constructed as groups whose members all had the same responsibilities. Decisions were reached through consensus, and conflicts resolved through discussion and compromise (Peters, 1994:118), and members were expected to contribute equally to the maintenance and running costs. At most boreholes today, in principle all costs are split into equal shares. An increasing number of syndicates have, however, relinquished this custom and members pay according to the number of cattle they keep at the borehole (interviews, four syndicate members in Kgatleng District, June 2000 and March 2007). The old system placed a heavy financial burden on those syndicate members who own the smallest herds, but at the same time it meant that the less wealthy members were regarded as equals.

The significant income diversity among syndicate members and the fact that many syndicate members have their source of income in the form of cattle instead of in wage payment becomes a problem when it is time to collect contributions. Often the time to pay contributions and the best time to sell an animal do not coincide. Although syndicate members are often tardy in paying their contribution, it is still a slow process for anyone to be punished. In extreme cases the inability to find cash can go on for years without the member being excluded. In some instances the tardiness is a protest against what is perceived as an unfair division of costs. Not all members who hold smaller herds want to contribute as much as the wealthier cattle holders. The situation is usually saved by the owners of the bigger herds. As they have a greater need for the borehole to function they are often forced to carry the costs of non-paying members. In some borehole syndicates, free-riding has become a major problem (interviews, one syndicate member and two private borehole owners in Kgatleng District, June 2000 and March 2007).

The unwillingness to punish, and in extreme cases exclude, members who do not fulfil their obligations towards the syndicate can be explained. Syndicate membership is usually passed from father to son and consequently the same families who created the syndicate are likely to be sharing it a few generations down the line. The Tswana tradition emphasises discussion, compromise and consensus rather than strict rules and punishment. This is especially the case with water resources, since accessing them is essential for the rural population's livelihood. Since the demand for rights to permanent water sources exceeds the availability there is little chance of ever acquiring membership in another syndicate and consequently exclusion is a very harsh punishment (interviews, three syndicate members and one private borehole owner, Kgatleng District, June 2000 and March 2007). However, despite significant lenience non-paying members are eventually excluded.

Apart from the owners, two other groups use both private and syndicate boreholes – hirers who pay for their water consumption and dependants who acquire their rights through kinship ties with the owners. The position of permanent hirer is usually granted to someone close to the owner, such as an old servant, a close friend or a relative, and it can be passed on through inheritance. For much the same reasons that syndicate members are indulgent towards each other, these individuals rarely risk exclusion. In the case of temporary hirers, though, the game is played primarily according to financial rather than social rules, and a temporary hirer who does not pay his dues is not likely to be welcomed back to the borehole (interview, one syndicate member in Kgatleng District, June 2000 and March 2007). Dependants do not have financial obligations towards the borehole owners, and they do not even control their own herds, as their animals are incorporated into their patron's herd. It is more or less impossible for a dependant to be excluded as a punishment for not abiding by the rules. Exclusion of dependants instead takes the form of the owner changing the constitution of the syndicate. There has been a general move towards decreasing the number of dependants (Peters, 1994; interview, one dependant in Kgatleng District, June 2000 and March 2007).

A large proportion of the animals watered at the boreholes today are strays (Oageng, 1999; interviews, two syndicate members in Kgatleng District, June 2000 and March 2007). The increasing livestock population and the decreasing distances between boreholes on the grazing range are factors that are worsening the problem. As boreholes are usually not fenced, it is in practice impossible to select and exclude certain animals that come to the water point. Watering strays is a matter of give and take, since all cattle owners are aware that some of their own animals may be drifting on the range and depend upon being watered at some other water point. This is especially true between neighbours who recognise each other's brands. To give water to thirsty animals is seen as a moral responsibility (interviews, one permanent hirer and two herd boys in Kgatleng District, June 2000 and March 2007).

To construct boreholes in the arable fields is a much easier legal process than drilling on the communal grazing range. These water points are supposed to be used for domestic purposes and watering smallstock while the larger livestock is expected to be kept away. There is no concern with spacing boreholes and land site applications can be obtained readily. Moving the cattle permanently into the arable fields seems to be becoming a common solution to the difficulty of accessing water points on the grazing range and it is likely that in the near future the clear spatial division between crop farming and cattle rearing activities will disappear (interviews, R Mojaphoko, Secretary, Sub-Land Board Registry, Artesia, and G Tlwane, 28 February 2007).

Within the institutional structure of the customary property rights, two parallel processes can be detected. There is a willingness to protect the poor and vulnerable by being lenient with free-riders, permanent hirers and dependants. In this way the negotiability within the customary property rights system is possibly slowing down the process of exclusion, as is claimed by Berry (2002), Odgaard (2003) and others. Meanwhile, negotiability is also part of a process of increased exclusion and polarisation of water resources, as Peters (2004) suggests. Non-paying syndicate members, dependants and temporary hirers are being excluded and their water rights are increasingly being amassed in the hands of a cattle and water elite, resulting in a more unequal distribution of water resources.

6. Concluding remarks

Botswana's cattle sector has not made a significant contribution to economic growth and development during the last four decades. Agro-pastoral smallholders in the communal areas, such as Kgatleng District, do not even produce at subsistence levels. This article claims that one important explanation for their low performance is the unequal distribution of water resources for agricultural purposes.

Smallholders in Botswana have experienced expropriation in the form of a polarisation process taking place at the local level within the institutional framework of prevailing customary property rights regimes. These regimes care for the basic needs of the population in general while at the same time allowing for water rights to be amassed in the hands of the cattle elite. The last 80 years show institutional continuity in property rights to boreholes and water resources. The conclusion is that there has been no political will, at national or district level, on the part of tribal leaders, the colonial administration or the independence government, either to safeguard the interests of the smallholders or to redistribute agricultural resources in favour of the smallholders. National structural inequality has gone hand in hand with local polarisation.

The agricultural policy and legislation from the 1920s to the present has been instrumental in polarising access to agricultural resources. It has demonstrably, however, not resulted in a productive cattle sector. This article suggests that it is time to try another strategy, a strategy of equity, including a redistribution of water resources held under customary law and tenure. Such a strategy could be instrumental in creating a productive smallholder sector that can play a part in future agricultural development in Botswana.

Acknowledgements

This article is part of the research project ‘The Role of Equity in Development’, funded by the Swedish International Development Cooperation Agency, Department for Research Cooperation (Sida/SAREC). The author acknowledges valuable comments from colleagues in her department and two anonymous referees.

Notes

¹BaKgatla is the Tswana sub-unit to which the majority of inhabitants of the present Kgatleng District belong.